7 Pitfalls in Building a Predictive Analytics Capability: a Zurich Case Study
Share this Session:
  Scott Lee   Scott Lee
  Conor Jensen   Conor Jensen
Analytic Operations Lead
Zurich NA


Tuesday, September 20, 2016
02:00 PM - 02:45 PM

Level:  Business/Strategic

One of the biggest challenges with Predictive Analytics (PA) is that it suffers from silver bullet syndrome. No one knows what it is, but everyone thinks it can do anything. So they are then therefore greatly offended when it fails – as it always will – to do everything. At Zurich NA, its Chief Data Scientist and PA Program Director, Conor Jensen worked with Knowledgent’s Analytics & Visualization Practice Lead, Scott Lee, to cure PA of that disease – at least within commercial product underwriting. And while the patient cannot quite yet be given a fully clean bill of health, she is certainly miles further from death’s door than before!

Presented with stark honesty and sometimes gallows humor, Jensen and Lee organize their talk around seven deadly pitfalls – choices or strategies that, if done poorly, will absolutely flatline your Predictive Analytics capability. But sidestepped and handled deftly, those same pitfalls can instead provide you with valuable experience aligning PA with your business’ needs.

The pitfalls have been drawn from across the whole gamut of capability challenges:

  • Technology platform
  • Data quality
  • Data politics; group “coöpetition”
  • Separation of concerns across business, data, and IT
  • Capability marketing / demand management
  • Governance and funding models
  • &c.

Scott Lee is a partner at Knowledgent, the Data and Analytics company. As its Analytics and Visualization lead, he builds data-fueled business services by mashing-up promising trade concepts, like enterprise information management (EIM) and data science, with well-established fields – industrial psychology, ergonomics, and human-computer interface (HCI). Mr. Lee works with clients worldwide on their most challenging data monetization, advanced analytics, and information value chain projects. His career spans 20 years of consulting experience with a focus on maximizing information value, especially in challenging climates such as post-merger integration and global IT/process change programs.

Conor Jensen is the Analytics Program Director and Lead Data Scientist for Zurich North America’s Predictive Analytics team. Along the way Conor has accumulated a wide-ranging skillset with an education in Pure Mathematics completely ignored while working as a retail store manager, military meteorologist, strategy consultant, actuary, and more. He is continuously looking for new problems to solve and new ways to cause trouble.

Close Window